[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Ya/RPpR3AdGAFtqX@sol.localdomain>
Date: Tue, 7 Dec 2021 13:25:18 -0800
From: Eric Biggers <ebiggers@...nel.org>
To: Matthew Wilcox <willy@...radead.org>
Cc: cgel.zte@...il.com, viro@...iv.linux.org.uk,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
Lv Ruyi <lv.ruyi@....com.cn>, Zeal Robot <zealci@....com.cn>
Subject: Re: [PATCH] fs/dcache: prevent repeated locking
On Tue, Dec 07, 2021 at 01:17:41PM +0000, Matthew Wilcox wrote:
> On Tue, Dec 07, 2021 at 10:16:46AM +0000, cgel.zte@...il.com wrote:
> > From: Lv Ruyi <lv.ruyi@....com.cn>
> >
> > Move the spin_lock above the restart to prevent to lock twice
> > when the code goto restart.
>
> This is madness.
>
> void d_prune_aliases(struct inode *inode)
> spin_lock(&inode->i_lock);
> if (likely(!dentry->d_lockref.count)) {
> __dentry_kill(dentry);
> goto restart;
> ...
> static void __dentry_kill(struct dentry *dentry)
> if (dentry->d_inode)
> dentry_unlink_inode(dentry);
> ...
> static void dentry_unlink_inode(struct dentry * dentry)
> spin_unlock(&inode->i_lock);
>
> Did you even test this patch?
This same wrong patch has been sent several times before. I think it's fair to
say that this code could use a comment, e.g.:
/* i_lock was dropped */
goto restart;
- Eric
Powered by blists - more mailing lists